Researchers from the German Research Center for Artificial Intelligence (DFKI) presented groundbreaking work at the 38th GI/ITG International Conference on Architectures of Computing Systems (ARCS 2025). Their paper, titled “Spend More to Save More (SM²): An Energy-Aware Implementation of Successive Halving for Sustainable Hyperparameter Optimization” introduces a novel approach that embeds energy-efficiency directly into the hyperparameter tuning process for machine learning models.
On June 3rd and 4th, 2025, all SustainML partners convened in person at eProsima’s office in Tres Cantos (Madrid) for a two-day workshop.
eProsima is proud to announce the release of SustainML v0.1.0, the first official version of their backend and developer framework for sustainable machine learning. A comprehensive backend library and end-to-end developer framework designed to bring environmental awareness to every stage of machine learning. This inaugural release delivers all the core plumbing for orchestrating tasks, querying metadata, profiling energy and latency, and shipping a rich, Docker-based front-end.
On April 25, 2025, INRIA, a prominent partner in the SustainML project, presented significant research findings at the CHI Conference on Human Factors in Computing Systems (CHI 2025). Their paper, titled "Should I Choose a Smaller Model?: Understanding ML Model Selection and Its Impact on Sustainability", delves deeply into the sustainability considerations within Machine Learning (ML) model selection processes, underscoring the pressing need for environmentally responsible computing practices.
eProsima is committed to fostering a more sustainable future. Leading the SustainML project, eProsima aims to significantly reduce the environmental impact of machine learning (ML) applications, reducing CO2 emissions and enhancing energy efficiency. By integrating the DDS communication layer into the SustainML framework, eProsima strives to optimize data sharing and enhance overall efficiency, aligning with SustainML’s ambitious environmental goals.
SustainML is among these nine innovative projects dedicated to creating a sustainable ML framework for Green AI.
Plaza de la Encina 10-11, Núcleo 4, 2ª Pl.
28760 Tres cantos - Madrid (España)
This project has received funding from the European Union’s Horizon Europe research and innovation programme under grant agreement No 101070408.
